A method for propagation estimation in XL MIMO networks

Abstract

The invention is related to a method for classifying propagation regions in extremely large multiple-input multiple-output (XL-MIMO) systems.

Description

[0001] DESCRIPTION

[0002] A METHOD FOR PROPAGATION ESTIMATION IN XL MIMO NETWORKS

[0003] Technical Field

[0004] The invention is related to a method for classifying propagation regions in extremely large multiple-input multiple-output (XL-MIMO) systems.

[0005] Prior Art

[0006] Accurately distinguishing between near-field and far-field propagation is a critical challenge for XL-MIMO systems. This distinction that is essential for optimizing beamforming, channel estimation, and resource allocation. Existing beamforming and channel estimation methods rely on simplified assumptions of either planar or spherical wavefronts, which fail to capture the complex signal behavior encountered in real-world scenarios, particularly with the increasing prevalence of mixed line-of-sight (LOS) and non-line-of-sight (NLOS) propagation conditions. These limitations lead to misaligned beams, inaccurate channel models, inefficient resource use, and degraded system performance, especially in applications such as mmWave / THz communications, and reconfigurable intelligent surfaces (RIS).

[0007] In prior art, propagation estimation methods in XL-MIMO systems primarily relied on simplistic wavefront models (planar or spherical) and predefined algorithms that lacked the flexibility to adapt to dynamic propagation environments.

[0008] Fixed threshold methods for near-field and far-field transition, [1,2,3] use a fixed threshold, such as the Rayleigh distance, to define the boundary between near-field and far-field regions. These methods fail to account for environmental factors, LoS / NLoS conditions, and multipath effects, leading to oversimplification and misclassification.

[0009] A planar wavefront assumption for far-field propagation [4] is widely used in classical multipleinput multiple-output systems (MIMO) and early extremely large multiple-input multipleoutput systems (XL-MIMO), this approach assumes uniform signal behavior across the array. However, this approach fails to capture the curvature of wavefronts in near-field scenarios, leading to significant errors in beamforming and resource allocation for users located close to the antenna array.A spherical wavefront models for near-field propagation [5] is employed in some advanced XL-MIMO and Reconfigurable Intelligent Surface (RIS) related systems, this approach assumes purely spherical wavefronts for near-field users. While this approach is accurate for near-field conditions, it breaks down in far-field regions or transitional areas, limiting its applicability in mixed-propagation environments.

[0010] Existing methods for propagation classification in XL-MIMO systems suffer from several limitations that hinder their effectiveness. First, they are inherently inflexible, relying on predefined approaches that fail to adapt to dynamic environmental conditions, such as transitions between line-of-sight (LoS) and non-line-of-sight (NLoS) propagation or varying multipath behaviors. Second, these methods perform poorly in transitional regions — the "grey areas" between near-field and far-field propagation — resulting in significant inaccuracies and degraded system performance. Finally, the misclassification of propagation types leads to inefficient resource utilization, causing suboptimal beamforming, channel estimation, and power allocation, particularly in dense deployment scenarios or high-mobility environments. As a result, all the problems mentioned above have made it necessary to provide a novelty in the related field.

[0011] Brief Description and Objects of the Invention

[0012] The main object of the present invention is to establish a method for accurately estimating the propagation characteristics in XL-MIMO systems by distinguishing between near-field and far-field propagation conditions, while also identifying and defining a transitional grey region where the classification between the two regimes becomes ambiguous.

[0013] Another object of the invention is to establish a method that enables more efficient beamforming, precise channel estimation, optimized resource allocation, and effective interference management.

[0014] Another object of the invention is to establish a method that enhances security within wireless communication systems. Accurate propagation estimation allows for better interference management and beam alignment, reducing the likelihood of eavesdropping or signal interception in sensitive communication channels.

[0015] To achieve above-mentioned advantages, the method comprises steps of establishing a communication link between a base station equipped with an XL antenna array and at least oneuser equipment (UE), wherein the base station is configured to measure signal characteristics; measuring a power profile of received signals over the XL antenna array to differentiate line-of-sight (LOS) and non-line-of-sight (NLOS) propagation conditions based on the spatial distribution of received power; measuring a delay profile of received signals to identify propagation regions by analyzing delay variation across the antenna array; determining a phase profile by converting the delay profile into phase information; calculating a distance profile from the delay profile to further classify near-field and far-field propagation; measuring a polarization profile to evaluate polarization behavior across the antenna elements and distinguish LOS and NLOS propagation scenarios; determining a Doppler profile from the received signals if mobility of the UE is detected, to analyze Doppler shifts across the antenna array; and aggregating the results from the power, delay, phase, distance, polarization, and Doppler profiles to classify the propagation region as near-field, far-field, or a mixed region, wherein the weights are adaptively assigned to each profile based on environmental conditions. Thus, the unique characteristics of signal propagation are captured across the antenna array by leveraging six distinct signal profiles which are power, delay, phase, distance, polarization, and Doppler profiles.

[0016] Description of the Figures of the Invention

[0017] The figures and related descriptions necessary for the subject matter of the invention to be understood better are given below.

[0018] Figure 1. A flow chart of the present invention.

[0019] Detailed Description of the Invention

[0020] The invention is related to a method for classifying propagation regions in extremely large multiple-input multiple-output (XL-MIMO) systems

[0021] For carrying out method, a communication link is established between a base station equipped with an XL array comprising of multiple antenna elements and a user equipped with a single or multiple antenna array having multiple antenna elements. The XL array at the base station can be arranged in various configurations, such as a uniform linear array (ULA), uniform planar array (UP A), or uniform circular array (UCA). The method described is applicable to any type of antenna array, as found in existing literature. The base station is capable of measuring thetime delay and received power strength. For example, through the Delay Profile, from the received signal transmitted by the user. Additionally, the base station can capture reflected signals from environmental obstacles.

[0022] Referring to Figure 1: after receiving the signal, the base station analyses the signal profiles.

[0023] For the power profile, the base station measures the received signal power over the XL antenna array. This measurement decides whether the signal received comes from LOS or NLOS.

[0024] A specific weight is assigned to each profile such that the sum of all weights equals wt.

[0025] Each profile provides a binary decision: one if the decision indicates Line-of-Sight (LoS) and zero for Non-Line-of-Sight (NLoS). The decision is then multiplied by the weight of the corresponding profile. For example, if the power profile determines based on its behavior that the signal is LoS, the weight of the power profile is multiplied by one. These weighted values are then summed to obtain wt. If the decision weight for being in LoS exceeds r / los, the final decision is classified as LoS. Conversely, if the total weight is less than r / los, the decision is classified as NLoS.

[0026] Each profile makes its binary decision based on its behavior. For instance, the power profile may exhibit a specific pattern where the power is maximized at certain antenna elements and minimized at others within the XL array. Based on this behavior, the power profile might determine that the signal is in LoS. The behavior of each profile can be analyzed using various algorithms and methods to improve the decision -making process. Additionally, Al techniques can be employed to enhance the binary decision of each profile.

[0027] One method for assigning weights can proceed as follows:

[0028] • If there is no Doppler effect, the power profile is assigned 50% weight (wl=0.5), while the delay and polarization profiles each receive 25% (w2=w3=0.25)

[0029] • If Doppler is present, the power profile is assigned 40% weight (wl=0.4) and the delay, Doppler, and polarization profiles each receive 20% (w2=w3=w4=0.2)

[0030] If the decision weight for being in LoS exceeds r / los, the final decision is classified as LoS. If the total weight is less than r / los, the decision is classified as NLoS. Various algorithms and methods can be employed to adapt the weight assignments for each profile.The threshold rjlos can be defined based on environmental factors or the accuracy of the XL transceiver design. For example, r / los could be set to 0.6.

[0031] In case of LoS decision of the first step; A specific weight y is assigned to each profile. This weight is determined based on the slope of the profile pattern. If the slope is equal to zero, the weight is also zero. By measuring the slope of the profile pattern, the weight of the profile is directly calculated, with the weight being equal to the slope.

[0032] The behavior and pattern of each profile can be studied using various algorithms and methods to improve the weighting process. Al techniques may also be employed to enhance the weighting of each profile.

[0033] The weight of each profile is compared against a specific threshold r / NF. If the weight exceeds the threshold, the decision for that profile is classified as near-field. Otherwise, the decision is classified as far-field forthat profile. The threshold r / NF can be defined based on environmental factors or the accuracy of the XL transceiver design. For example, r / NF might be set to 0.6.

[0034] If all profiles yield a near-field decision, the final decision is classified as near-field. If all profiles yield a far-field decision, the final decision is classified as far-field. However, if at least one profile's decision differs from the others, the final decision falls into the gray zone.

[0035] The final decision-making process can be further improved by applying Al techniques.

[0036] In the case of a Non-Line-of-Sight (NLoS) decision in the first step, a specific weight cr is assigned to each profile. This weight is determined based on the multiple slopes of the profile pattern. If the slope is equal to zero, the weight is also zero. By measuring these slopes, appropriate weights can be assigned.

[0037] The behavior and pattern of each profile can be analyzed using various algorithms and methods to improve the weighting process. Additionally, Al techniques can be applied to enhance the accuracy of the weighting for each profile.

[0038] The weight of each profile is then compared against a specific threshold r / NF. If the weight exceeds the threshold, the decision for that profile is classified as near-field. Otherwise, the decision is classified as far-field for that specific profile.The threshold r / NF can be defined based on environmental conditions or the accuracy of the XL transceiver design. For instance, r / NF might be set to 0.6.

[0039] If all profiles yield a near-field decision, the final decision is classified as near-field. If all profiles yield a far-field decision, the final decision is classified as far-field. However, if at least one profile's decision differs from the others, the final decision falls into the gray zone.

[0040] For a LOS scenario, the power distribution across the array will exhibit a continuous and predictable pattern, concentrated in a single VR. This is due to the strong direct LOS path dominating the signal. Here, the profile has a single smooth transition in the power across the array. In this case, LOS detection is achieved. Large weight value is given to this profile.

[0041] For a NLOS scenario, the power profile is likely fragmented into multiple, sparse visibility regions (VR). This occurs because the signal reaches different portions of the array through distinct paths (reflections, diffractions, or scattering), and some array elements may not be illuminated by the user due to blockages. In this case, it is unable to detect whether the signal is from NLOS near-field propagation or far-field propagation. Less weight is given to this profile.

[0042] For the power profile, the base station measures the delay of the received signal across the XL array. This measurement has different behaviors in LOS and NLOS scenarios.

[0043] In a LOS scenario, the user equipment can be modeled as a single dominant cluster, resulting in a delay profile characterized by one primary component (corresponding to the direct path) and smaller secondary components arising from the curvature of the spherical wavefronts. In this case, this profile is given high weight similar to the power profile.

[0044] In a NLOS scenario, the delay profile becomes more complex, typically featuring multiple clusters. Each cluster corresponds to a distinct multipath component (e.g., reflections or diffractions), with a main delay component for each cluster surrounded by secondary components due to spherical wavefronts at the cluster level. The pattern of delay variation across the array decides whether this variation comes from NLOS in near-field propagation or far-field propagation. The near-field propagation gives smooth delay transition over the antennaarray. In this case, the delay profile is given high weight for near-field far-field estimation decision.

[0045] The effects of spherical wavefronts are more pronounced at higher bandwidths, as wider bandwidths provide finer resolution in delay and better capture the curvature of the wavefronts.

[0046] A phase profile can be calculated from the delay profile by converting the time into phase as <p = — — T, where T is the delay profile and is the wavelength. Since this profile is similar to the delay profile, equal weight is given here with similar decisions as the delay profile.

[0047] The distance profile can be calculated from the delay profile as r = TC, where c is the speed of light in free space. Since this profile is similar to the delay profile, equal weight is given here with similar decisions as the delay profile.

[0048] A polarization profile is also measured, polarization rotation can occur due to the Faraday effect or anisotropies in the medium, but these effects are generally small for short paths or ideal environments. Using polarized antennas, the polarization profile is measured over the antenna array. This measurement has different behaviors in LOS and NLOS.

[0049] In near-field LOS, the polarization shows some periodic behavior considering the polarization over the first / reference element a0. Given that the polarization over the antenna elements 2TT

[0050] follows an= a0+ Aan, where Aan= — Arn, this behavior can be easily noted over the array and hence the polarization profile is given high weight for estimation decision.

[0051] In far-field LOS, the polarization angle would be equal to the transmitted one and equal distributed over the antenna elements. If this behavior is noted, it is assumed that the signal is propagated in the far-field region. Here, the polarization profile is given high weight for estimation decision.

[0052] In NLOS, the signal reaches the array after interacting with multiple obstacles (e.g., walls, buildings, or other scatterers). Each interaction (reflection, diffraction, or scattering) can alter the polarization of the wave. Consequently, the received signal at different antenna elements can exhibit varying polarization states. These variations depend on: The nature of the scattering surfaces (e.g., smooth, rough, or metallic), and the angle of incidence and the geometry of the scattering paths. While individual multipath components may retain consistent polarizationwithin their paths, the combined signal arriving at the array typically exhibits depolarization (a mix of polarization states).

[0053] In far-field NLOS, the polarization angle received over the array is a mixture of polarizations received from different paths; however, the effect of this combination is assumed to be equal / similar over all antenna elements. Here, the polarization profile is given high weight for estimation decision.

[0054] In near-field NLOS, the polarization angle received over the array is a mixture of polarizations received from different paths and since the polarization over the antenna elements in near-field follows specific distribution (an= a0+ Aan),the polarization profile can be used to distinguish the near-field and far-field propagation. Here, the polarization profile is given high weight for estimation decision.

[0055] A Doppler profile can be measured if the user is in mobility, the Doppler profile is measured by extracting the phase of consecutive symbols in the received signal (0), then Doppler profile is measured as fDis the rate of phase change.

[0056]

[0057] In far-field propagation, the Doppler profile is fixed over all antenna elements in both LOS and NLOS. Therefore, if the measurement of this profile is fixed over the antenna array, the propagation is conserved to be far-field propagation. Here, the Doppler profile is given high weight for near-field far-field decision.

[0058] In near-field propagation, the measurement exhibits some linear / nonlinear variation over the antenna array regarding it comes from LOS or NLOS. The slopes of these variations can be measured and tracked to calculate whether the signal comes from near-field, far-field, moving from near-field to far-field, or moving from far-field to near-field. Here, the Doppler profile is given high weight for near-field far-field decision.

[0059] For classification decision, the results from all profiles are aggregated to classify the propagation region. There are three possible outcomes of the classifying. First, all profiles agree on the near field, the signal is detected as near-field. Second, all profiles agree on the far-field, the signal is detected as far-field. Third, the result from the profiles are mixed (at least one profile results differs than other), the signal is detected as gray area (mixed area).In a preferred embodiment, Al mechanism can be cooperated to adapt the weight of each profile according to the environment changing for better propagation detection.REFERENCES

[0060] [1] G. Liu and X. Sun, "Two-Stage Matrix Differencing Algorithm for Mixed Far-Field and Near-Field Sources Classification and Localization," in IEEE Sensors Journal, vol. 14, no. 6, pp. 1957-1965, June 2014, doi: 10.1109 / JSEN.2014.2307060.

[0061] [2] A. M. Molaei, B. Zakeri and S. M. Hosseini Andargoli, "Components Separation Algorithm for Localization and Classification of Mixed Near-Field and Far-Field Sources in Multipath Propagation," in IEEE Transactions on Signal Processing, vol. 68, pp. 404-419, 2020, doi: 10.1109 / TSP.2019.2961226.

[0062] [3] H. Yan, Y. Wang, Y. Gong, Z. Zhang and L. Wang, "Improved Sparse Symmetric Arrays Design for Mixed Near-Field and Far-Field Source Localization," in IEEE Transactions on Aerospace and Electronic Systems, vol. 59, no. 6, pp. 7486-7498, Dec. 2023, doi: 10.1109 / T AES.2023.3291678.

[0063] [4] Lu, L., Li, G.Y., Swindlehurst, A.L., Ashikhmin, A. and Zhang, R., 2014. An overview of massive MIMO: Benefits and challenges. IEEE journal of selected topics in signal processing, 8(5), pp.742-758.

[0064] [5] Lu, H., Zeng, Y., You, C., Han, Y., Zhang, J., Wang, Z., Dong, Z., Jin, S., Wang, C.X., Jiang, T. and You, X., 2024. A tutorial on near-field XL-MIMO communications towards 6G. IEEE Communications Surveys & Tutorials.

Claims

CLAIMS1. A method for classifying near-field and far-field propagation in extremely large multiple-input multiple-output systems, characterized bya. establishing a communication link between a base station equipped with an XL antenna array and at least one user equipment (UE), wherein the base station is configured to measure signal characteristics;b. measuring a power profile of received signals over the XL antenna array to differentiate line-of-sight (LOS) and non-line-of-sight (NLOS) propagation conditions based on the spatial distribution of received power;c. measuring a delay profile of received signals to identify propagation regions by analyzing delay variation across the antenna array;d. measuring a polarization profile to evaluate polarization behavior across the antenna elements and distinguish LOS and NLOS propagation scenarios;e. determining a Doppler profile from the received signals if mobility of the UE is detected, to analyze Doppler shifts across the antenna array; andf. aggregating the results from the power, delay, phase, distance, polarization, and Doppler profiles to classify the propagation region as near-field, far-field, or a mixed region, wherein weights are adaptively assigned to each profile based on environmental conditions.

2. A method according to Claim 1, wherein the XL antenna array is configured as one of a uniform linear array (ULA), uniform planar array (UPA), or uniform circular array (UCA).

3. A method according to Claim 1, wherein LOS conditions are identified when the power profile exhibits a continuous and predictable distribution concentrated within a single visibility region (VR).

4. A method according to Claim 1, wherein the delay profile distinguishes near-field and far-field propagation based on the smoothness of delay transitions across the XL antenna array.

5. A method according to Claim 1, wherein the polarization distinguishes near-field and far-field propagation by detecting specific polarization distribution patterns across the antenna elements.

6. A method according to Claim 1, wherein the Doppler profile is analyzed to identify transitions between near-field and far-field propagation based on variations in Doppler shift across the array.

7. A method according to Claim 6, wherein the Doppler profile is measured as fD=the rate of phase change.

8. A method according to Claim 1 , characterized by further comprising step of determining a phase profile by converting the delay profile into phase information.

9. A method according to Claim 1, characterized by further comprising step of calculating a distance profile from the delay profile to further classify near-field and far-field propagation.

10. A method according to Claim 1, characterized by further comprising step of using of an artificial intelligence (Al) mechanism to dynamically adjust weights assigned to each profile based on environmental factors.

11. A method according to Claim 1, wherein the distance profile is calculated using the relationship r = TC, where c is the speed of light.2.TTC 12. A method according to Claim 1, wherein the phase profile is determined as <p = — A —. T, where T is the delay profile and is the wavelength.

13. The method of claim 1, further comprising a decision-making step where the total weight and profile-specific weights are used to classify the propagation region into: • near-field propagation if all profiles consistently indicate near-field,• far-field propagation if all profiles consistently indicate far-field, ora gray zone if conflicting classifications are observed across the profiles.

14. A data processing system comprises means for carrying out the method according to any of the preceding claims.

15. A computer program comprising instructions which, when the program is executed by a data processing system of claim 14, cause the computer to carry out the method according to Claim 1 to 13.

16. A computer-readable data carrier having stored thereon the computer program of claim 15.

Images